3.2 Data Time period:
This research includes informations from KSE100 index, Interest rate ( KIBOR rates ) , Exchange rate and Inflation Rate of 7 old ages on monthly footing. The period ranges from June, 2005 to August, 2012.
Four variables are used in the research, we will happen the consequence of Interest rates, Exchange rate and Inflation Rate on return on stocks ( KSE 100 Index ) , so:
3.3.1. Interest rate ( KIBOR rates ) , Exchange rate and Inflation Rate as Independent Variable.
3.3.2 KSE 100 index as Dependant Variable.
3.3.1. Interest Rates ( KIBOR ) :
KIBOR means Karachi interbank offer rates which are defined as lending/borrowing rates quoted by the Bankss. The Bankss under this agreement quote these rates at specified clip i.e. 11.30 AM at Reuters. Presently 20 Bankss are member of KIBOR nine and by excepting 4 upper and 4 lower extremes, rates are averaged out that are quoted for both terminals ie offer every bit good command. The quotation mark rates available in KIBOR scopes from one hebdomad to 3 old ages. KIBOR is used as a benchmark for corporate loaning rates.
Interbank Rate in Pakistan reduced to 9.19 % in Jan of 2013 from 9.28 % in Dec of 2012. Interbank Amount in Pakistan is reported by the State Bank of Pakistan. Traditionally, from 1991 until 2013, Pakistan Interbank Rate averaged 10.50 Percent accomplishing an all clip high of 17.42 Percent in May of 1997 and a record depression of 1.21 Percentage in July, 2003.In Pakistan, the interbank rate is the rate of involvement charged on short-run loans made between Bankss. This figure below shows historical informations for Pakistan Interbank Rate.
Beginning: Pakistan Bureau Of statistics
3.3.2 KSE 100 Index:
Karachi Stock exchange 100 index is stock index moving as a criterion to compare monetary values on Karachi Stock exchange over a period of clip. The company with highest market capitalisation is selected. In this index high market capitalisation from each sector is besides included. The index was launched in Nov.1991 with basal point of 100 points. By 2001, it had grown to 1770 points and reached 12,285 in Feb 2007.A twenty-four hours before Internet Explorer 26 December 2007, former Prime Minister Benazir Bhutto was assassinated, it was on record high of all time 14,814 points. During Global Crisis 2008 it decreased to 9,187 points. It reached to enter highest degree on November 7, 2012 which is 16,218 points which is now taken as emerging market in Asia. This figure below shows historical informations for Pakistan Karachi Stock Exchange.
Beginning: Karachi stock Exchange
3.3.3 Exchange Rate:
Exchange rate is the value of one currency in footings of another currency. Exchange rate may positively or negatively affect stock return depending on the economic system of the state.
Historically, from 1988 until 2012, the USDPKR averaged 58.91 making an all clip high of 98.11 in December of 2012 and a record depression of 4 in May of 2010. It is calculated on day-to-day footing. This figure below shows historical informations for Pakistan Exchange rate.
Beginning: State Bank of Pakistan
3.3.4 Inflation Rate:
Inflation is increase in monetary values of goods and services due to which people will purchase smaller sum of goods with the same sum of money. Inflation negatively affects stock returns because net incomes of the houses ‘ lessenings with addition in monetary value of goods due to increased costs.
In Pakistan, Inflation Rate was recorded at 7.9 per centum at the terminal of the twelvemonth 2012. It is reported by Pakistan Bureau of Statistics. This figure below shows historical informations for Pakistan Inflation Rate.
Beginning: Pakistan Bureau Of statistics
3.4 Statistical Tool Used in Research
3.4.1 Multiple Regression Model:
Multiple arrested development analysis is a statistical tool for understanding the relationship and their impact between two or more variables. It involves two variables the dependant variable which is to be explained and independent variable which is the extra explanatory variables that are thought to bring forth or be associated with alterations in the dependant variable. Normally, it has two or more independent variables.
3.4.2 Model of the Research:
To look into the effects of macroeconomics variables on Stock Returns we use a Multiple Regression Model ” .
R= KSE 100 Index.
ITR= Interest Rate ( KIBOR ) .
IFR= Inflation Rate.
ER= Exchange Rate.
3.4.3 Components of the Multiple Regression Model:
188.8.131.52 Coefficient of Determination or R-square ( R2 ) :
It is statistic value which shows the anticipation of future results on footing of other given information. Normally, its value ranges from 0 to 1.0. R2 value if it is closer to 1 it means that information is fits the arrested development line good and frailty versa. It measures the per centum of the fluctuation in the dependant variable produced by independent variables. R-square is the most normally used step of goodness-of-fit of a arrested development theoretical account.
184.108.40.206 Standard Error of the Coefficient ; Standard Error ( Se ) :
It is the step of the fluctuation of a parametric quantity estimation or coefficient about the true parametric quantity. The standard mistake is a standard divergence that is calculated from the chance distribution of estimated parametric quantities.
220.127.116.11 Statistical Significance:
A trial used to cipher the grade of association between a dependent variable and one or more independent variables. If the deliberate p-value is smaller than 5 % , the consequence is said to be statistically important ( at the 5 % degree ) . If P is greater than 5 % , the consequence is statistically undistinguished ( at the 5 % degree ) .
A trial statistic that describes how far an estimation of a parametric quantity ( it is a characteristic of population ) is from its hypothesized value ( i.e. , given a void hypothesis ) .If a t-statistic is sufficiently big ( in absolute magnitude ) , an expert can reject the void hypothesis.
It is besides known as deliberate chance. P-value is the estimated chance of rejecting the void hypothesis. The larger the p-value, the more likely the void hypothesis is true.
3.5 Problems in Regression Model
The undermentioned jobs may originate in arrested development theoretical account. These jobs should be removed in order to do the arrested development theoretical account more perfect:
Autocorrelation is besides sometimes called lagged correlativity ” or consecutive correlativity ” , which refers to the correlativity between members of a series of Numberss arranged in clip. Positive autocorrelation might be considered a specific signifier of continuity ” , a inclination for a system to stay in the same province from one observation to the following.
Autocorrelation is normally found in time-series informations. Time-series informations are normally homoscedastic in nature.
It can be found in both transverse sectional informations and clip series informations. In transverse sectional informations mold, informations drawn from one part may reflect the features of the adjacent parts, it is known as spacial car correlativity.
Autocorrelation in economic clip series is a contemplation of the civilization and institutional traditions of the populations which produced during the series, In other words, what people did in the yesteryear would impact their present and future. In short, Autocorrelation would happen if the theoretical account is non right specified.
18.104.22.168 Effect of Presence of Autocorrelation in Data
The presence of autocorrelation does non do prejudice in the appraisal of the theoretical account coefficients, but it reduces the efficiency of a theoretical account for prediction because it increases the discrepancy of the remainders every bit good as the discrepancy of the estimated carbon monoxide efficient. As they both are reciprocally related to each other so increase in discrepancy will cut down the efficiency of the theoretical account.
22.214.171.124 Test for the Detection of Autocorrelation:
There are several statistical trials available for observing the autocorrelation in a theoretical account. Most frequently, the undermentioned two trials are used:
The Ocular Test ( Residual Plot ) :
Residual is usually defined as the difference between the existent and predicated values of dependent variables. The standard mistake of the estimation is the standard divergence of the remainders
A residuary secret plan is a graph that depicts the residuary values on perpendicular axis and the independent variable on horizontal axis. If the points in secret plan of remainders are indiscriminately dispersed around horizontal axis or it does non exhibit any systematic order or any pattern so there is autocorrelation in the theoretical account.
The Durban Watson Test:
Most calculating package like E-Views, Gretel and SPSS calculate the value of Durban Watson Test automatically but Microsoft Excel does non. The d-test is more powerful for theoretical accounts based on big samples than little samples. Some writers suggest that every bit long as D- trial value is less than 2.5 and greater than 1.5 void hypothesis should accepted. In other words, the theoretical account used is free from Autocorrelation. Normally, the figure of observations should more than 30.
3.5.2 Heteroscedasticity Test:
Hetero ” means unequal and scedasticity ” agencies spread ( discrepancy ) so the word Heteroskedasticity is the unequal distribution of remainders. As from the secret plan of the remainders it is non in the systematic mode, it is plotted in unsystematic mode. The trial shows that there is no heteroscedasticity in the information. The antonym of heteroscedasticity is homoskedasticity.
Heteroscedasticity arises in volatile high-frequency time-series informations such as day-to-day observations in fiscal markets and in cross-section informations where the graduated table of the dependant variable and the explanatory power of the theoretical account tend to change across observations. Microeconomic informations such as outgo studies are typical. The perturbations are still assumed to be uncorrelated across observations.
In calculating mold, an inefficient theoretical account would hold larger prognosis mistakes than an efficient theoretical account, therefore it is non good to hold heteroscedasticity in the theoretical account. It can seen easy seen signifier the distribution of remainders.
126.96.36.199 Causes of Heteroscedasticity:
There are many causes of heteroscedasticity. The following are likely most common:
Where database of incorporating big value and the other incorporating little value i.e. the scope between the smallest and the largest value is really big.
Where the grade of growing rates between the dependant variable and independent variable vary significantly. It is largely common in clip series informations mold.
It besides occurs where informations is heterogenous. E.g. income degree informations vary significantly from people to people, & A ; at response to certain merchandise, high income positions will different from that of low income degree.
188.8.131.52 Effect of Heteroscedasticity
The chief impacts of heteroscedasticity are
It does presence does non do the coefficient estimations biased but it causes the discrepancies to increase of OLS estimations to increase. It means that in perennial samplings, the estimated carbon monoxide efficient will fluctuate more widely than they usually do.
Its presence causes the underestimate of the discrepancies of the coefficients. This could annul the T and f trial which will misdirect the modellers to reject the void hypothesis, when it should be accepted.
As the discrepancies increases, the efficiency of the theoretical accounts lessenings. So the prognosis of the theoretical account would non be accurate.
In any arrested development theoretical account it is good to hold unbiased estimations but besides efficient estimations. Therefore, it is non good to hold heteroscedasticity in a theoretical account.
184.108.40.206 Test for the Detection of Heteroscedasticity:
The Ocular Test ( Residual Plot ) :
A residuary secret plan is a graph that depicts the residuary values on perpendicular axis and the independent variable on horizontal axis. If the points in secret plan of remainders are indiscriminately dispersed around horizontal axis or it does non exhibit any systematic order or any pattern so there is Heteroscedasticity in the theoretical account.
Multicollinearity occurs when two or more forecasters in the theoretical account are correlated. In other words, when there is an exact or about exact additive relation among the independent variables.
Multicollinearity is non a job if the end is merely predict Y from a set of variables. The overall consequence will still be accurate and R2 quantifies how good the theoretical account predicts the Y value. It becomes a large job if the end is to happen impact of set of independent variables ( X ) on dependant variable ( Y ) . Then two jobs would originate:
First, p-values would be misdirecting. It means that a p-value will high even if the variable is of import.
Second, the assurance intervals on the arrested development coefficients will be really broad. The assurance intervals may even include zero, which means one ca n’t even be confident whether an addition in the X value is associated with an addition, or a lessening, in Y. Because the assurance intervals are so broad, excepting a topic ( or adding a new one ) can alter the coefficients dramatically and may even alter their marks.
220.127.116.11 Beginnings of Multicollinearity
There are four primary beginnings of Multicollinearity:
The informations aggregation method employed
Constraints on the theoretical account or in the population.
An over defined theoretical account.
3.6 Research hypothesis:
H0: Macroeconomic indexs affect Stock Returns
H1: Macroeconomic indexs affect do non Stock Returns.
3.7 Theoretical model
The theoretical model of the survey consist of dependent variable and independent variable
KSE 100 Index
Interest Ratess ( KIBOR Ratess )